Machine-learning–guided design of nanotherapeutics for personalized treatment of uterine fibroids

Over the past years, nanotechnology has become a promising approach to delivering drugs more effectively, whereas machine learning methods offer great solutions of optimizing the design of the treatment, depending on the unique aspects of a patient. Uterine fibroids belong to the category of the most frequent benign tumor in women of childbearing age and may cause serious clinical issues as pain in the area of the pelvis, irregular bleeding, and infertility. Traditional methods of treatment are not very personalized and can cause either poor therapeutic outcomes or unwanted side effects. This paper has shown a machine-learning-based design and assessment of nanotherapeutics to achieve the customization of the treatment of uterine fibroids. A sample population comprising of 420 cases was created that contained clinical and therapeutic important variables such as the age of patients, body mass index (BMI), fibroid size, severity of symptoms, type of nanoparticles, dose of drug, and ligand of targeting. These were inputs to be used in predicting treatment efficacy which is the expected level of effectiveness of the nanotherapeutic intervention. To assess the associations between patient properties, nanocarrier properties, and the therapeutic outcomes expected, both statistical analysis and graphic visualisation were used to investigate the relationships. The findings show that patient related variables as well as the parameters used in the design of nanotherapeutics play a major role in determining the performance of the treatment. Specifically, predicted treatment efficacy is enhanced by the optimization of the dosage of drugs as well as by the choice of targeting ligands. The results indicate the possibility of combining machine learning and nanomedicine to allow individualized treatment plans to increase precision of treatment and better clinical results in patients with uterine fibroids.